I can see that mypaint recognises not only pressure but width of touch. Maybe we could use that info for fake multitouch (a small touch area that expands => pinch, a wide touch area that moves => two finger drag)

Also, I don't know how deep access we have to the hardware, but my understanding is that the resistive screen gives out an x,y for the touch

If I touch two points simultaneously eg 0,5 and 7,10 the wires that get shorted would be the 0 and 7 on the x axis and the 5 and 10 on the y axis. So the hardware can't possibly distinguish between my touch and the equivalent of 0,10 and 7,5.

So we can't move two balls arbitrarily on the screen with this info, but the info is absolutely enough for pinch, rotate, two finger drag etc...

multitouch on synaptics touchpad for example is just one set of coordinates and a seperate variable stating the number of fingers.

Also multitouch on the nexus one and other htc phones is not full (check out android app multitouch pong, it clearly states which devices and why) By the way, with the two x, and two y's (e.g. the four point set) even multitouch pong could be done on the N900. The 'fakeness' of the implementation would be baringly noticeable.

Finally maybe the pressure sensitivity/area data could be used for user interaction alternatively: heavy press and light press accounting for different interactions
________
Live sex